GreenChain — AI Supply Chain Carbon Auditor

GreenChain is an AMD-powered multi-agent AI system designed to solve one of the biggest challenges in corporate sustainability: supply chain carbon tracking. Companies are failing their ESG goals because they cannot see the emissions hidden in their tier-2 supplier networks. Our system changes that. Built for the AMD Hackathon 2026 by Redcore Collective, GreenChain uses three specialized AI agents working in sequence. Agent 1 (The Data Extractor) parses raw shipping manifests and CSV files to extract structured shipment records. Agent 2 (The Carbon Calculator) applies IPCC-aligned emission factors to calculate CO₂ per route and identifies the single biggest carbon hotspot in the supply chain. Agent 3 (The Sourcing Strategist) searches a database of verified eco-friendly suppliers and generates a prioritized Sustainability Roadmap with estimated CO₂ savings per recommendation. The system is powered by Llama-3-70B running on AMD Instinct MI300X via ROCm, which provides the high memory bandwidth needed to process large, complex supply chain datasets faster than traditional GPU solutions. The frontend is built with Streamlit and includes a Demo Mode so judges can explore the platform instantly without API credentials. Key features include real-time carbon footprint analysis, automatic carbon hotspot detection, green supplier recommendations via semantic search, downloadable audit reports in markdown, and support for any global shipping route using live geocoding. The project is fully open source and deployed on Streamlit Cloud.